An iOS app that reads my own training data and answers questions about it: whether pace is drifting away from heart rate, how this week's load compares to the ramp, whether the fatigue in my notes shows up in the numbers.

How it works: Strava activities and Apple Health metrics land in a Postgres-backed Spring Boot service → the Flutter client streams answers from a coach that reasons over that history instead of giving generic fitness advice.

Built with: Flutter Dart Riverpod Java 21 Spring Boot 3 PostgreSQL Flyway OpenAI Gemini

The engineering I'm actually practicing here:

• Two repos joined by a machine-checked contract — the client vendors the backend's generated OpenAPI spec, and a test suite binds every operation to the Dart that calls it and the Dart that parses the response. A backend change that would break the shipped app turns the client's CI red.
• A compatibility verdict, not a diff — each spec change is classified by what it does to a client already in the field (breaking / restricting / additive), and a shape the rules don't model fails the gate rather than falling through to a reassuring answer.
• Ports and adapters — Strava ingestion, HealthKit sync, and two LLM providers each sit behind their own boundary.
• Streaming chat over SSE, with cancellable in-flight responses and persisted conversations.

Both repos are private while this is in development.

A full-stack webapp that tailors your resume to any job description and scores it against real ATS systems.

How it works: Pick a job from the built-in job board or paste any job description → AI analyzes keyword gaps, role alignment, and formatting.

🔗 Try it live (Chrome-only): re-zoo-me.com

Built with: Next.js TypeScript FastAPI Python MongoDB PostgreSQL Redis

Key features:

• Block-based resume editor with drag-to-reorder
• Multi-stage ATS scoring pipeline
• One-click PDF export via WeasyPrint